UDSKRIFT 🔗
WEBVTT
00:02.460 --> 00:11.442
[SPEAKER_01]: You could only do so much in a test environment and a lot of the way that we had trained the model was from test data.
00:11.462 --> 00:16.323
[SPEAKER_01]: And it'd be probably at a huge offshore team that we'd have tons and tons of falls.
00:16.343 --> 00:21.005
[SPEAKER_01]: We had an outsourced company that also was able to reduce tons and tons of test data.
00:21.065 --> 00:22.205
[SPEAKER_01]: We actually had our own,
00:22.805 --> 00:26.087
[SPEAKER_01]: proprietary data sets that we train them all on as well.
00:26.608 --> 00:33.272
[SPEAKER_01]: But there really is nothing quite like field data where data that's like in real environments.
00:34.193 --> 00:37.175
[SPEAKER_01]: My name is Mike Link and the COO of
00:44.073 --> 00:47.115
[SPEAKER_02]: A podcast bringing you interviews with tech visionaries.
00:47.195 --> 00:48.416
[SPEAKER_02]: Six months moonlighting goes.
00:48.436 --> 00:49.937
[SPEAKER_00]: It's the last and last of the backends.
00:50.398 --> 00:52.859
[SPEAKER_02]: Who share what it takes to change an industry?
00:53.000 --> 00:54.521
[SPEAKER_00]: I don't exactly know what to do.
00:54.541 --> 00:55.982
[SPEAKER_02]: The main goes to get right.
00:56.842 --> 01:00.785
[SPEAKER_02]: Who built the teams that have their bad company is its people.
01:00.805 --> 01:03.487
[SPEAKER_03]: The teams help each other achieve this proud of our team.
01:03.667 --> 01:05.929
[SPEAKER_02]: Keeping scalability top of mind.
01:06.269 --> 01:07.530
[SPEAKER_02]: All that infrastructure was up there.
01:07.550 --> 01:09.011
[SPEAKER_02]: Yes, they've been fighting it as we grow.
01:09.131 --> 01:10.012
[SPEAKER_01]: Total waste of time.
01:10.492 --> 01:12.693
[SPEAKER_02]: The stories you don't read in the headlines.
01:13.193 --> 01:14.613
[SPEAKER_03]: It's not an easy thing to achieve.
01:14.673 --> 01:16.014
[SPEAKER_02]: To yourself, it does it often.
01:16.234 --> 01:16.834
[SPEAKER_02]: Try to begin.
01:17.454 --> 01:20.535
[SPEAKER_02]: To ride the ups and downs of the startup line.
01:20.635 --> 01:22.215
[SPEAKER_03]: Need to really want it.
01:22.295 --> 01:23.576
[SPEAKER_02]: Not just about technology.
01:23.636 --> 01:24.276
[SPEAKER_02]: All this and more.
01:24.916 --> 01:25.796
[SPEAKER_02]: On code story.
01:27.417 --> 01:29.338
[SPEAKER_02]: I'm your host, Noel Apphart.
01:29.358 --> 01:34.119
[SPEAKER_02]: Today, how Mike Leak is driving the spread of detection and response.
01:34.839 --> 01:37.480
[SPEAKER_02]: To the epidemic of senior citizen falls.
01:44.257 --> 01:50.163
[SPEAKER_02]: Mike Link grew up in Rochester, New York and had no intention of getting into technology, but he loves it now.
01:50.644 --> 01:55.489
[SPEAKER_02]: His educational background was a neuroscience, as he found the brain absolutely fascinating.
01:56.190 --> 02:02.256
[SPEAKER_02]: As he looked into Western medicine he figured out that he wasn't that attractive to him, nor the lifestyle that he wanted.
02:03.103 --> 02:07.446
[SPEAKER_02]: He got into building his own things, including a trip to Indian returning into sales.
02:08.186 --> 02:13.049
[SPEAKER_02]: Outside of his professional life, he lives in San Diego and loves travel, visiting Mexico City often.
02:13.349 --> 02:19.053
[SPEAKER_02]: He's a huge fan of cooking as well, and when he travels, he enjoys experiencing different cultures to the actual people living there.
02:22.226 --> 02:26.529
[SPEAKER_02]: Mike was found by his current venture as he had passed experience in Elder Care.
02:27.289 --> 02:33.273
[SPEAKER_02]: He joined the team to help with the go-to-market planning, sales, et cetera, to be the Swiss Army Knife at the company.
02:34.374 --> 02:41.058
[SPEAKER_02]: When they started digging into senior care and the problems around fall detection, he led the group and how they took on this market.
02:44.100 --> 02:47.162
[SPEAKER_02]: This is Mike's creation story at Kami Vision.
02:52.199 --> 02:54.321
[SPEAKER_01]: Camille Vision is an AI vision company.
02:54.742 --> 03:01.128
[SPEAKER_01]: We're probably going to be making a transition from just AI vision to AI vision and AI agents.
03:01.348 --> 03:02.570
[SPEAKER_01]: We soon will provide.
03:03.110 --> 03:06.814
[SPEAKER_01]: But we started in 2014 in home secure.
03:06.974 --> 03:10.157
[SPEAKER_01]: This is when the home security boom was happening.
03:10.918 --> 03:15.283
[SPEAKER_01]: The founder of CEO had been a PhD in the chip guy.
03:16.003 --> 03:33.920
[SPEAKER_01]: and figured out how to make these chips that could do AI on the edge and home security cameras that he knew the manufacturer for, so he was like, okay, let's go out at this home security business that blew up and between then and 2022 and high-journed
03:34.420 --> 03:38.665
[SPEAKER_01]: The business got pretty big and became a funding source for other projects.
03:39.266 --> 03:46.975
[SPEAKER_01]: One of the things that you've been working on is an enterprise security product, and then the other thing that I came on when I was the first starting at Campion.
03:46.995 --> 03:52.942
[SPEAKER_01]: Fortunately, too, was bringing and helping to get the fall detection product ready.
03:53.522 --> 03:59.425
[SPEAKER_01]: I came on because they found me and they knew that I had experience in elder care and senior care.
04:00.065 --> 04:07.269
[SPEAKER_01]: So they said, okay, why don't you come on and help us understand how do we sell this and how do we go to the market and what makes sense and probably seem models all these things.
04:08.610 --> 04:12.992
[SPEAKER_01]: And over the years, I've helped build the elder care product.
04:13.192 --> 04:20.516
[SPEAKER_01]: and then now helping clean up the home product and helping with the enterprise security product as well.
04:20.596 --> 04:33.984
[SPEAKER_01]: So I do it all and different products are the home security, but you could think like ring, visit, bring city tea, a camera
Load full transcript
and a home multiple cameras in a home doorbell cans, security and safety package action dog detection.
04:34.744 --> 04:40.186
[SPEAKER_01]: And then we have the older care product, which started in 2022.
04:40.926 --> 04:54.509
[SPEAKER_01]: And this was born out of both requests and unsolved niche in senior care, where people had home that had home security cameras to say, hey, we have a cloud recording of Mom and Dad's fallen.
04:54.549 --> 04:55.230
[SPEAKER_01]: We found them.
04:55.810 --> 04:58.671
[SPEAKER_01]: We looked back at the cloud footage and we saw that they had fallen.
04:58.771 --> 05:00.791
[SPEAKER_01]: But we never got notified.
05:04.292 --> 05:06.033
[SPEAKER_01]: mom or dad was on the ground already.
05:06.693 --> 05:21.857
[SPEAKER_01]: This is interesting and then when they had looked into the senior care market, a lot of the technology was very expensive from both the CapX and a monthly recurring standpoint, and this is because a lot of the tech was built with VC.
05:22.977 --> 05:28.064
[SPEAKER_01]: Money, big teams, expensive software, they went to all the conferences.
05:28.804 --> 05:32.989
[SPEAKER_01]: So the price of the product actually had to be high, they had no other choice.
05:33.450 --> 05:37.775
[SPEAKER_01]: Whereas the cami, the reason I was drawn to the group was because there was no outside funding.
05:37.815 --> 05:38.316
[SPEAKER_01]: It was just...
05:38.856 --> 05:41.877
[SPEAKER_01]: all from within the business, which meant that we can be very low cost.
05:42.478 --> 05:49.701
[SPEAKER_01]: We had really good camera manufacturers, so we could have very low cost hardware at on top of the low cost software.
05:50.361 --> 05:53.102
[SPEAKER_01]: And the team was very nimble on agile.
05:53.182 --> 05:56.204
[SPEAKER_01]: If there was an idea that I had, I was like, okay, we should do this.
05:56.764 --> 05:59.505
[SPEAKER_01]: We should do this type of detection and we should have set up this way.
06:00.206 --> 06:03.327
[SPEAKER_01]: This is how I got involved with the candy care product.
06:03.407 --> 06:22.176
[SPEAKER_01]: And jumped into the other thing, then the final thing, and this is the thing I went to Europe to work on a bit was the enterprise product, and we're probably going to go out and retail with this, but intelligence over existing devices and retail, for both security and for business intelligence, which is going to be maxed for the poll that we're going to be jumping into.
06:22.976 --> 06:24.877
[SPEAKER_01]: The biggest base of it is the home product.
06:24.897 --> 06:26.698
[SPEAKER_01]: There's millions of cameras all look at the world.
06:27.378 --> 06:33.500
[SPEAKER_01]: The second would be the elder care product where we have primarily customers in senior care on the US.
06:33.961 --> 06:40.223
[SPEAKER_01]: We have an influx of senior care facilities in UK and then also in Japan.
06:41.364 --> 06:51.288
[SPEAKER_01]: And we have a bit in consumer as well where probably
about 500 to 1000 families have the devices in consumer home in the US.
06:53.326 --> 07:02.589
[SPEAKER_02]: Mike, let's dig into the first project or initiative that you came into, can we vision to take on, right?
07:02.609 --> 07:04.490
[SPEAKER_02]: Now talk about MVP's on this show a lot.
07:04.610 --> 07:07.871
[SPEAKER_02]: What was your MVP, the first thing that you took on?
07:08.491 --> 07:11.712
[SPEAKER_02]: When you joined the company, how long did it take you to go after?
07:11.812 --> 07:14.713
[SPEAKER_02]: And what sort of tools or methodologies you're using to attack it?
07:16.575 --> 07:19.776
[SPEAKER_01]: the MVP was the fall detection product.
07:20.416 --> 07:26.138
[SPEAKER_01]: It was basically ready, and I gave recommendations on okay, maybe we should tweak it this way that way.
07:26.638 --> 07:32.600
[SPEAKER_01]: The model had already been built by the team, and they really needed somebody to go to market.
07:33.180 --> 07:37.004
[SPEAKER_01]: So the MVP was there, but it was really broken in a way.
07:37.084 --> 07:42.849
[SPEAKER_01]: So I think they had gone to the MVP very quickly, but in a lot of ways it was flawed.
07:43.009 --> 07:50.797
[SPEAKER_01]: I think it wasn't necessarily about making the MVP, but it was how do you get the MVP to appoint the sticks?
07:52.218 --> 08:00.106
[SPEAKER_01]: You could only do so much in a test environment and a lot of the way that we had trained the model was from test data.
08:01.027 --> 08:06.113
[SPEAKER_01]: And it'd be probably at a huge offshore team that we'd have tons and tons of falls.
08:06.173 --> 08:08.275
[SPEAKER_01]: We had an outsourced company that also,
08:08.515 --> 08:10.796
[SPEAKER_01]: We was able to produce tons and tons of test data.
08:10.836 --> 08:15.897
[SPEAKER_01]: We actually had our own proprietary data sets that we were able to train the model on as well.
08:16.398 --> 08:23.040
[SPEAKER_01]: But there really is nothing quite like field data, where data that's like in real environments.
08:23.600 --> 08:28.342
[SPEAKER_01]: And I think that was probably the biggest hurdle for us was getting people to use it.
08:29.002 --> 08:36.484
[SPEAKER_01]: And then using the misses or the mistakes as data points to train the model.
08:38.366 --> 08:41.037
[SPEAKER_01]: Looks like this was missed, but this was barely caught.
08:41.548 --> 08:42.809
[SPEAKER_01]: and now we can go back and train it.
08:42.889 --> 08:49.873
[SPEAKER_01]: And part of the problem is like a pretension and early on, we had a lot of churn and customers because we were figuring it out.
08:50.573 --> 09:02.800
[SPEAKER_01]: And I think it was a delicate balance where it was like, okay, you're trying to do a million things and then find new customers in a space where the sales cycle is fairly long.
09:03.740 --> 09:06.562
[SPEAKER_01]: And yeah, it was definitely a grind.
09:06.862 --> 09:07.963
[SPEAKER_01]: You'd find a customer
09:08.627 --> 09:31.832
[SPEAKER_01]: feed on board and when sales you have to have them up those confidence product and then they start and then maybe there's some mistakes and then you've got to pair them for these mistakes and then train the models to help fix the mistakes and then give a lot of resources there was a lot of burn to cash but I would say that that's didn't make the product better and after a while we finally got the product in a stable place probably after a year
09:32.472 --> 09:56.140
[SPEAKER_01]: I would say 8 months to a year, where we weren't seeing the same bugs, we weren't seeing the issues that the devices were having, we weren't seeing as many misses, and then we're able to make claims on real data, probably over 40 facilities or something like this initially that we had gotten, and done thousands of detections, follow detections with, because that was the main part of us follow detections.
09:58.242 --> 10:03.103
[SPEAKER_02]: Then the next thing I'm interested in is if to wrap that wrap this question in a box it's around roadmap.
10:03.243 --> 10:11.645
[SPEAKER_02]: I'm curious about how you progress in mature things, how you have progress in mature things, at can be visioned in for the products and the portfolio.
10:12.325 --> 10:15.686
[SPEAKER_02]: And then again, to wrap that in a box a little bit when looking for it, how do you build your roadmap?
10:15.726 --> 10:16.706
[SPEAKER_02]: How do you go about deciding that?
10:16.766 --> 10:19.667
[SPEAKER_02]: Okay, this is the next most important thing to build or to address.
10:21.310 --> 10:42.848
[SPEAKER_01]: The first most important thing to build in a dress It always comes from the user or the use case And I think the same thing with the road map I always feel like And it's like a cliche thing Or you hear all these founders or CEOs say this like The best products are always customer-led And like even meta-open AI and traffic like all the products are the
10:43.194 --> 10:45.416
[SPEAKER_01]: the customers that we are the products basically.
10:45.496 --> 11:00.787
[SPEAKER_01]: So I think the product really came from the need in the market where there's so many unwitness falls, and it's such a big pain point, and you might have a mom or a dad, Noah, who's fallen at home,
11:01.587 --> 11:28.930
[SPEAKER_01]: Maybe they've sat there for three or four hours and you find out and it's just a heart-wrenching feeling And I think this is where the first thing came from And I think in selling and in speaking to people who were in the market either at home or In a facility setting That's when I discovered with the next step to where the product needs to go and I realized that The fall is one piece
11:29.811 --> 11:35.334
[SPEAKER_01]: and often the perception is that it's too late once you've detected a fall.
11:35.895 --> 11:41.498
[SPEAKER_01]: And it partially is true because once you've detected a fall, this is the sign of decline, right?
11:42.058 --> 11:47.922
[SPEAKER_01]: Maybe decline started in
January, and Mama Dad just fell now in its September.
11:48.882 --> 11:55.888
[SPEAKER_01]: So the fall is the sign of decline and a lot of people want to know how do we prevent that fall for happening?
11:55.908 --> 11:58.290
[SPEAKER_01]: How do we prevent falls not just to text falls?
11:58.931 --> 12:06.297
[SPEAKER_01]: And one of the things that we found was that people typically don't hurt themselves the first or the second or the third time that they fall.
12:06.637 --> 12:19.011
[SPEAKER_01]: It's typically by like the sixth or the seventh or the eighth time that they need to go to the hospital because they broke a hip where they have a fracture or where they got a concussion or something like this, right?
12:19.632 --> 12:26.239
[SPEAKER_01]: So if we can capture the early falls, we have a greater chance preventing the serious falls.
12:26.968 --> 12:51.027
[SPEAKER_01]: not the only caveat to that is that even some of the early falls, it's still a bit late and this is speaking from like the customer mind where it's like, we will pay for all the tech right but there has to be something else that we could do to prevent the fall and there's so many other data points that can be collected now with modern technology that you could actually
12:54.765 --> 12:58.748
[SPEAKER_01]: kind of slight decline from happening because you have a pulse on somebody's health.
12:58.768 --> 13:06.715
[SPEAKER_01]: The reason a lot of these folks either at home or in facilities or at a climb to climb is because they only got a doctor as so many times a week.
13:07.216 --> 13:09.197
[SPEAKER_01]: They're only getting checked on so many times a week.
13:10.209 --> 13:17.454
[SPEAKER_01]: and there's not probably not enough implementation to lifestyle activities that can help prolong.
13:18.055 --> 13:32.485
[SPEAKER_01]: So I think having these conversations with people and realizing that fall prevention is like the buzzword that sells, but people were really looking for a way to live longer in place.
13:32.865 --> 13:34.687
[SPEAKER_01]: And the word is aging in place, but
13:35.427 --> 13:36.848
[SPEAKER_01]: This was the real problem.
13:37.008 --> 13:45.070
[SPEAKER_01]: The fall is like the sewer or like the blemish, but at the core, it's the aging and place thing.
13:45.270 --> 13:46.771
[SPEAKER_01]: The people don't want to leave home.
13:47.151 --> 13:48.712
[SPEAKER_01]: They don't want to go to a facility.
13:49.212 --> 13:50.652
[SPEAKER_01]: They don't want to lose independence.
13:51.373 --> 14:00.656
[SPEAKER_01]: And the way that we saw getting there was basically taking all the data that we can capture using modern tools and technology.
14:01.336 --> 14:06.359
[SPEAKER_01]: to make inferences on health and predict basically what's going to happen in the future.
14:06.860 --> 14:15.965
[SPEAKER_01]: There's tons and tons of peer-reviewed journals on fall risk and what types of medications and comorbidities cause fall risk.
14:16.386 --> 14:29.634
[SPEAKER_01]: There's also tons of peer-reviewed journals on different changes in behaviors and in vitals that can lead to not just fall risk, but other health risks or negative health outcomes.
14:30.094 --> 14:52.464
[SPEAKER_01]: And I think what we would like to do and what we started working on basically is taking these data points and how do we build a system that incorporates not just fall detection, but we can take a look at the morbidities, medications, lifestyle, and then make recommendations for real time as if a doctor or therapist.
14:53.144 --> 14:57.645
[SPEAKER_01]: or caregiver was able to be there every morning, right, remotely.
14:57.945 --> 14:59.445
[SPEAKER_01]: And nobody has to go and bother this.
14:59.865 --> 15:04.686
[SPEAKER_01]: I don't have to call Aunt Suzie or Grandma Jane and say, hey, what's going on?
15:04.726 --> 15:07.307
[SPEAKER_01]: Did you take your blood pressure this morning?
15:07.327 --> 15:08.247
[SPEAKER_01]: Did you do this?
15:08.407 --> 15:09.467
[SPEAKER_01]: Did you go for a walk?
15:10.007 --> 15:14.748
[SPEAKER_01]: As an adult child, I'll be for the look on my device and CO.
15:14.848 --> 15:17.909
[SPEAKER_01]: It looks like mom has it been sleeping much lately, right?
15:17.949 --> 15:18.909
[SPEAKER_01]: Maybe I should check out her.
15:18.929 --> 15:19.329
[SPEAKER_01]: It looks like
15:19.989 --> 15:23.191
[SPEAKER_01]: activity has declined in the last four weeks.
15:23.752 --> 15:24.972
[SPEAKER_01]: She's moved less and less.
15:25.653 --> 15:28.895
[SPEAKER_01]: Now, I get in a notification that says, hey, I'm as less active.
15:28.935 --> 15:29.896
[SPEAKER_01]: Please check on her.
15:30.016 --> 15:31.076
[SPEAKER_01]: This is a risk, right?
15:31.537 --> 15:39.442
[SPEAKER_01]: So it's like all these little data points that actually allow someone to be healthier longer, but it's not just about the data, it's about the implementation.
15:39.462 --> 15:46.026
[SPEAKER_01]: I think that's a next challenge because once you have the data, once you can give the information to the individual, now they actually have to act on it.
15:46.390 --> 15:49.112
[SPEAKER_01]: Alright, so will the family be able to do this at home?
15:49.612 --> 15:51.373
[SPEAKER_01]: Will the care staff be able to do this?
15:52.013 --> 15:56.736
[SPEAKER_01]: I think that's a future problem to solve, but gathering the data I think is the first step.
15:58.557 --> 16:00.278
[SPEAKER_02]: I'm curious about team, right?
16:00.318 --> 16:07.702
[SPEAKER_02]: How do you go about building your team and what you look for in those people to indicate that they're the winning horses to join you?
16:09.577 --> 16:33.229
[SPEAKER_01]: I think what I have looked for has been non-traditional experience and it's been more so an overall attitude and excitement and eagerness and curiosity in somebody's ability to go out of their scope or maybe be a critical thinker in the way that they prop them solve.
16:33.909 -->
16:35.070
[SPEAKER_01]: And it's really hard to
16:36.210 --> 16:41.032
[SPEAKER_01]: sus this out in a brief conversation or an interview or a brief chat.
16:41.192 --> 16:51.236
[SPEAKER_01]: I think boys being referred somebody is great but I always say the excellent people like exceptional people always have good opportunities for the most part already.
16:51.836 --> 17:00.439
[SPEAKER_01]: I think I've gotten pretty lucky that it's just so happened that I've known some really exceptional people that just were happening to look for work.
17:02.874 --> 17:09.158
[SPEAKER_01]: knew what I was working on and knew me and then we'd have a conversation and they were eager to work on this thing with me.
17:09.839 --> 17:31.034
[SPEAKER_01]: But yeah, I kind of feel like curiosity and like ultimate accountability were the things that I look for and looked for with some of the team members that I added meaning, you know, and a lot of the time I'll do like, hey, we're gonna do some work for a while, maybe even on a consulting basis and then if we both feel like it's a fit, you can come on
17:32.695 --> 17:38.036
[SPEAKER_01]: a higher because I kind of feel like the best way to assess out somebody's work is to have the work.
17:38.236 --> 17:45.198
[SPEAKER_01]: People can tell you like a million beautiful things and show up and have a million excuses why the problem doesn't get solved.
17:45.298 --> 17:57.941
[SPEAKER_01]: But I think it's like the curiosity to ask questions like even if I have a sales guy or a marketing girl or somebody like it's like the second and third and fourth level questions that bring them to
17:58.881 --> 18:01.903
[SPEAKER_01]: solution and then having ultimate accountability.
18:02.043 --> 18:08.588
[SPEAKER_01]: If you have the curiosity to ask the questions about whatever the problem is, whether it's, we have a poor domain score in.
18:09.248 --> 18:12.030
[SPEAKER_01]: How does that work and then how is one solve that?
18:12.130 --> 18:14.532
[SPEAKER_01]: How do you open subdoppings so that you avoid this?
18:15.032 --> 18:25.502
[SPEAKER_01]: how do we clean the list and that it's like sometimes somebody would just stop although we have a bad domain score like what do we do or how do we fix this or we have this problem and sales outreach.
18:25.562 --> 18:29.346
[SPEAKER_01]: We have this problem where nobody wants to buy at this point or this guy is saying this thing, right?
18:29.406 --> 18:29.687
[SPEAKER_01]: It's like
18:30.367 --> 18:55.593
[SPEAKER_01]: somebody who kind of asked why would he say this and then you think about this and you're like okay it's not that he has a problem with a specific term that we have in the agreement it's that he doesn't necessarily believe it the part does what we say it does so does he need a red line or does he just need to be given more information and ask questions like it's some of these things that that's hard to figure out and interview but you can very quickly figure it
out when
18:57.840 --> 19:00.381
[SPEAKER_01]: I think also, like, there's a lot of change that happens.
19:00.501 --> 19:05.143
[SPEAKER_01]: There's a lot of people that have come and go at thing in the last four or five years.
19:05.563 --> 19:17.368
[SPEAKER_01]: And I think not being afraid to let somebody go pretty quickly when it's not working out because a lot of times you can see either the red flag to the signs that this just isn't a fit.
19:17.429 --> 19:23.151
[SPEAKER_01]: And you can give someone all the time in the world, but as soon as you get that feeling like you know, that you gotta get somebody different.
19:27.250 --> 19:30.075
[SPEAKER_02]: Okay, I'm curious about scale, right?
19:30.156 --> 19:36.868
[SPEAKER_02]: How are you approaching scale with camera vision and have there been interesting years where you've had to fight it as you've grown?
19:37.069 --> 19:39.674
[SPEAKER_02]: Scale could be people, it could be technology, it could be all the things.
19:41.040 --> 19:52.986
[SPEAKER_01]: Initially, it was me and a very small team, when I was the primary one-bravenant business, and then was probably like a year and a half past, and we're like, okay, we're gonna get a sales team, right?
19:53.566 --> 19:56.728
[SPEAKER_01]: Had like maybe 10 sellers, and...
19:57.708 --> 19:59.390
[SPEAKER_01]: It was just not working.
19:59.630 --> 20:00.791
[SPEAKER_01]: I was doing something different.
20:01.151 --> 20:06.096
[SPEAKER_01]: There was a sales leader who wasn't necessarily being the best sales leader.
20:06.196 --> 20:13.163
[SPEAKER_01]: Like we'd have demos and meetings and pipeline, but none of it was really real and I could tell that the sellers were saying it.
20:14.404 --> 20:16.826
[SPEAKER_01]: It was just a not a good situation, a lot of the overhead.
20:17.867 --> 20:41.397
[SPEAKER_01]: for the people and I basically got rid of the whole sales team and we kind of scaled back and it actually helped us move faster in terms of like scaling the business because we moved away from selling directly and just sort of going to partners and said, why do I hire sales team when I can go and find partners that will give this to their existing customers and they already have their relationship, but I just need to give them a fee.
20:42.157 --> 20:56.426
[SPEAKER_01]: So that's the direction we moved in, and we've kept the team super slim, mainly using AI for a lot of things, design and email automation and some of the operational stuff in the back and it
20:59.227 --> 21:04.811
[SPEAKER_01]: Vibed, coded, and then worked with engineers to build a dashboard database.
21:05.891 --> 21:08.053
[SPEAKER_01]: And that's been how we've remained.
21:08.313 --> 21:14.176
[SPEAKER_01]: Slamma, I think we probably will scale in the next year or two of the team a bit.
21:14.676 --> 21:24.983
[SPEAKER_01]: But I think with a lot of the AI tools now, it's not required to have the same density of team that you might have needed five years
ago.
21:25.871 --> 21:46.110
[SPEAKER_01]: And the other thing, in terms of scale with the product is that as you get more and more users, especially in such a sensitive space, like fall detection, nursing homes, even people at home, you really have to have support out of them because you gotta think about it, like with 500 people, 500 people use in the service.
21:46.730 --> 21:49.593
[SPEAKER_01]: There's probably not tons and tons of falls on a daily basis, but
21:50.034 --> 21:59.074
[SPEAKER_01]: With thousands of people using it, there's lots of falls detected and you need to be able to respond to either a request or a, you know.
21:59.733 --> 22:13.017
[SPEAKER_01]: concern very quickly because trust is paramount for these people in for the relationship with them because if somebody's on the ground and they don't get notified or they can't see what's going on and they can't see what happened then.
22:13.898 --> 22:19.880
[SPEAKER_01]: So a really bad day, especially if they're injured they need to review the clip of the fall of biomedical purposes.
22:19.940 --> 22:26.722
[SPEAKER_01]: So I think support's been a key thing that we had to make sure that we had a place at basically now have a very good
22:27.343 --> 22:34.106
[SPEAKER_01]: outsourced team that can respond 24-7 to request that come in and help with technical troubleshooting.
22:34.646 --> 22:39.369
[SPEAKER_01]: And yeah, there's the most part are really good with getting to customer quickly.
22:41.109 --> 22:46.912
[SPEAKER_02]: So as you step out on the balcony, look across all that you have achieved with Camille Vision and what you've built.
22:47.252 --> 22:48.033
[SPEAKER_02]: What do you most proud of?
22:49.396 --> 22:54.319
[SPEAKER_01]: I think I'm most proud of having brought nothing to something.
22:55.079 --> 22:56.100
[SPEAKER_01]: I think that's the thing.
22:56.160 --> 23:04.424
[SPEAKER_01]: I was thinking of the day I was like, man, we started with no customers, no marketing, no product, website was very bare bones, all the things.
23:04.724 --> 23:07.566
[SPEAKER_01]: And I think I'm always impressed at how far we made it.
23:07.586 --> 23:12.308
[SPEAKER_01]: There was definitely a lot of times along the way where we were just like, man, I think this is it.
23:12.328 --> 23:13.429
[SPEAKER_01]: I think we're gonna shut down.
23:13.449 --> 23:18.792
[SPEAKER_01]: I kind of even forgot about that until you just asked the question, just now, but yeah, there was a lot of times where I started,
23:19.700 --> 23:40.044
[SPEAKER_01]: getting consulting gigs or I was like, huh, this might be the last month in Kami and then would get some great customers or get a big customer, get a big contract, for them to go keep on the lights or keep the the motion going and I guess proud of the overcoming dark times, so to say.
23:42.285 --> 23:42.585
[SPEAKER_02]: Excellent.
23:42.645 --> 23:44.385
[SPEAKER_02]: Okay, let's flip the script a little bit.
23:44.405 --> 23:47.886
[SPEAKER_02]: Tell me about a mistake you made and how you and your team responded to it.
23:49.657 --> 24:04.457
[SPEAKER_01]: Initially when we were having new hardware shipments come in it was just like a need thing like we needed the hardware as fast as we can get it and we brush to get the hardware because we'd sell everything that we had before.
24:05.398 --> 24:28.538
[SPEAKER_01]: And I remember one time we rush shipped, I think it was like a 1,000 or maybe 2500 devices, and we didn't have the manufacturer do any type of testing or anything or proof of QA or anything like this, and basically all of the devices ended up having
24:29.399 --> 24:52.154
[SPEAKER_01]: faulty cables, and we basically had to go through and replace all the cables and get new cables from the manufacturer, but in hindsight, it would have been as simple as getting the manufacturer to do or show us the QA so that we could have prevented this from happening, but it really stopped.
24:52.454 --> 24:55.256
[SPEAKER_01]: It probably took us like a month to get enough cables.
24:55.296 --> 24:56.377
[SPEAKER_01]: Initially we were just buying one
24:57.838 --> 25:14.602
[SPEAKER_01]: we had from a supplier and but it was a nightmare to get a up ship and like that and have almost 7,800 cameras that we had to ship out in every single one had a cable that did work and in fact was making like the device overheat a bit.
25:14.702 --> 25:22.984
[SPEAKER_01]: Thankfully we were able to get new cables that work were safe very quickly and but yeah it was just especially on top of the
25:28.165 --> 25:39.190
[SPEAKER_01]: 90 days, and this is when there was some challenges, so I think it was like a little bit over 90 days that it took to arrive and we tried to rush it, but that's just what happened.
25:39.330 --> 25:52.135
[SPEAKER_01]: And you kind of got a balance with people who have signed contracts and they're waiting for hardware, waiting for people at the reason you kind of got to manage that relationship carefully and have them not to lose patience on you, but yeah, that was a big oopsie.
25:55.652 --> 25:56.954
[SPEAKER_02]: Okay, let's move forward then.
25:56.994 --> 25:59.558
[SPEAKER_02]: What does the future look like for CamuVision?
25:59.618 --> 26:02.022
[SPEAKER_02]: For all that you've built for all that you've achieved there?
26:02.342 --> 26:05.407
[SPEAKER_02]: Where the industry is going, there's lots of change in technology.
26:05.427 --> 26:06.328
[SPEAKER_02]: What does the future look like?
26:07.607 --> 26:26.201
[SPEAKER_01]: Yeah, the future at Kami Vision looks like a shift from a lot of physical AI and when cameras and we do a little bit of robotics versus company doing this AI vision thing to more non physical AI agents.
26:26.941 --> 26:30.603
[SPEAKER_01]: a software that can run over other software.
26:31.204 --> 26:43.892
[SPEAKER_01]: I think specifically we are looking to turn can we care the elder care business into a much more comprehensive
aging and play solution like I described.
26:44.552 --> 26:48.215
[SPEAKER_01]: We're planning to bring the technology to the consumer market.
26:48.635 --> 27:13.373
[SPEAKER_01]: Later this year, early 2027, and ideally we could start to accumulate these different data points, like I said, when become a bit more integrated into the medical record, where maybe we will use the comorbidities and maybe medicationless and collect vitals to give people at home a better idea of whether health is at.
27:13.813 --> 27:16.835
[SPEAKER_01]: so they can know when decline is coming so they can stay safer.
27:17.415 --> 27:25.299
[SPEAKER_01]: The other thing will be the same thing in facilities and communities so that staff can provide better care in facilities and communities.
27:25.819 --> 27:30.042
[SPEAKER_01]: And then I think the project that I had mentioned earlier, the enterprise security product.
27:30.802 --> 27:47.190
[SPEAKER_01]: This thing I think is a natural progression from the home security cameras where we could basically give insights to retailers or outfit a commercial real estate site to do AI detections and make it a bit custom.
27:47.570 --> 27:50.112
[SPEAKER_01]: A lot of the solutions out there really are cookie cutter.
27:50.132 --> 27:56.355
[SPEAKER_01]: This is of course like enterprise customizations, but with the capabilities of AI now,
27:56.895 --> 27:59.177
[SPEAKER_01]: there's a lot more customization that you could do.
27:59.618 --> 28:09.908
[SPEAKER_01]: And especially us being pretty agile, we see a niche where we could basically allow customers to start to build their own detections, partnering with us.
28:10.128 --> 28:19.918
[SPEAKER_01]: Those are a few of the things that I see in that the near future, primarily the first step being the scaling and the growth of the elder care.
28:20.518 --> 28:31.501
[SPEAKER_01]: Business, not just in the States, but also in Europe, also in Japan, potentially also we might be going into Singapore, but yeah, actively looking for partners globally for that business.
28:32.921 --> 28:34.062
[SPEAKER_02]: Awesome, Mike, let's switch to you.
28:34.422 --> 28:36.102
[SPEAKER_02]: Who influences the way that you work?
28:36.442 --> 28:39.483
[SPEAKER_02]: I'm a person or many persons or something, you look up to and why.
28:40.730 --> 28:44.613
[SPEAKER_01]: Primarily have been the people that I have worked for.
28:44.633 --> 28:59.646
[SPEAKER_01]: This guy named Tom Bang, who initially hired me, is brilliant guy, I live in San Diego not far away from me, and I think I took a lot from him early on and learning, like, and honestly a lot of it's like mirroring and mimicry.
28:59.786 --> 29:02.588
[SPEAKER_01]: Like, I was always impressed by some of the people that I worked for.
29:03.369 --> 29:15.294
[SPEAKER_01]: in the way that they spoke, and the way that they went about things, the questions they asked, how they carried themselves, and how prepared they were, or how knowledgeable they were, and certain subjects.
29:16.054 --> 29:28.099
[SPEAKER_01]: And I think it would inspire me to go back to the drawing, water, go back to my desk, and figure that out, I'd be like, oh, what are they talking about, or I'd ask them, and then I'd go study.
29:28.299 --> 29:31.120
[SPEAKER_01]: And I think that was, those were probably my biggest influences, Tom,
29:33.241 --> 29:34.462
[SPEAKER_01]: where we're very closely with.
29:34.763 --> 29:49.597
[SPEAKER_01]: And at this point, feels like a bottom line, Felix Donchew, I probably talked to him more than any girlfriend or even my mom in the span of a year and a half when we were together, we'd be on the phone like most of the day doing things.
29:50.438 --> 29:54.202
[SPEAKER_01]: And just seeing how we moved and operating in the world, I think it was a big,
29:54.960 --> 30:13.528
[SPEAKER_01]: like inspiration, but also I picked up so many things on how to get things done quickly and how to manage people and finance and legal and sales and marketing and advertising and all the things that she's run, a ton of companies and yeah tons of other people but I think like typically it's the people that I see.
30:14.408 --> 30:23.070
[SPEAKER_01]: doing lots of different things that I find interesting and a lot of times it's, you know, people that I find like, almost a bit elusive.
30:23.090 --> 30:24.810
[SPEAKER_01]: I'm like, how do they do these things?
30:24.850 --> 30:26.450
[SPEAKER_01]: Like, these things don't even make sense to me.
30:26.490 --> 30:30.551
[SPEAKER_01]: But they're moving at such a pace and they're getting so many things done.
30:30.651 --> 30:32.812
[SPEAKER_01]: I feel like, okay, this is, I should pick this up.
30:33.532 --> 30:38.153
[SPEAKER_01]: But yeah, I would say that's kind of the influence on how I were.
30:38.553 --> 30:39.873
[SPEAKER_01]: I also think like early on,
30:40.827 --> 30:57.855
[SPEAKER_01]: both my parents worked super super hard and they always emphasized hard work to my grandfather and they didn't really come from anything I didn't really come from anything but work ethic was always really valued so I think that was like ingrained in me and I found that
30:58.461 --> 31:16.656
[SPEAKER_01]: the more volume I did, the better I got, and whatever I did, seeing that and having that influence and even the people that I worked for, having that influence and then that behavior being ingrained in me through practice, I think that was the biggest influence.
31:18.467 --> 31:23.572
[SPEAKER_02]: Last question, so you get on a plane, you're sitting next to a young entrepreneur who's built the next big thing.
31:24.052 --> 31:24.893
[SPEAKER_02]: They're jazzed about it.
31:24.933 --> 31:27.495
[SPEAKER_02]: Can't we show off the world and can we show off the you right there in the plane?
31:28.236 --> 31:30.718
[SPEAKER_02]: What advice do you give that person having gone down this road a bit?
31:32.315 --> 31:35.799
[SPEAKER_01]:
get as many people they use it as fast as possible.
31:36.720 --> 31:38.922
[SPEAKER_01]: Once it's built, it'll be already have it built.
31:38.962 --> 31:42.927
[SPEAKER_01]: And even if it's not built, see how many people you can get to agree to use it.
31:43.147 --> 31:51.236
[SPEAKER_01]: Once it is in a final form or an MPP form, because until you have people using it and until people
31:52.597 --> 31:55.259
[SPEAKER_01]: in real life, use the thing that you built.
31:55.879 --> 31:58.521
[SPEAKER_01]: You really don't know what you don't know.
31:59.141 --> 32:09.248
[SPEAKER_01]: And there's honestly so many challenges I feel like that come with first users or first use case even when I've built things for myself outside of the chemical thing.
32:09.488 --> 32:19.975
[SPEAKER_01]: It's amazing I'll have an app or a tool or something that I've coded using the coding tools or vibe code or something like this and how much Q&A that I'll have to do with myself
32:20.695 --> 32:21.696
[SPEAKER_01]: What did I want it to do?
32:21.736 --> 32:23.617
[SPEAKER_01]: And I'm like, oh, this doesn't really do this thing.
32:23.637 --> 32:27.240
[SPEAKER_01]: And then I have to go back and switch something around and I'm like, oh, I missed this thing.
32:27.481 --> 32:28.261
[SPEAKER_01]: Yeah, so I don't know.
32:28.301 --> 32:34.006
[SPEAKER_01]: I feel like having initial users is really helpful because it takes you.
32:34.626 --> 32:37.869
[SPEAKER_01]: It takes you outside of yourself because you feel like this thing is cool.
32:38.309 --> 32:41.432
[SPEAKER_01]: But other people need to feel like this thing is cool and then have
32:42.052 --> 32:47.217
[SPEAKER_01]: a continuous use, they should come back to this thing that you've built or this tool.
32:47.938 --> 32:54.525
[SPEAKER_01]: And then you'll have so many questions from them, I'm sure, though, ask about this thing, or ask about this thing.
32:55.286 --> 33:05.156
[SPEAKER_01]: The other thing is that the people who use your thing early on should feel very comfortable giving you like negative feedback.
33:05.696 --> 33:32.244
[SPEAKER_01]: I think a lot of times sometimes people tell you that the thing is amazing or the thing is while congrats you built this thing but the most valuable types of customers are the ones that will tell you that your product sucks and that it really isn't that great and that there's difficult things about it because then you can reverse engineer these these problems that they have and make it much better but yeah I always I always you know my answer that question is always users it's the best stick that's many people that use it because
33:32.984 --> 33:42.930
[SPEAKER_01]: That's your best early success to show to potential team members who want to join you and help build this thing or investors or other companies sell.
33:43.270 --> 33:46.712
[SPEAKER_02]: Love all that Mike, that's fantastic advice pointing back to the users.
33:47.172 --> 33:51.475
[SPEAKER_02]: Mike, thanks for being on the show today and thank you for telling your creation story at CammyVish.
33:52.035 --> 33:52.896
[SPEAKER_01]: Yeah, thank you so much.
33:52.956 --> 33:53.656
[SPEAKER_01]: Now I was great.
33:56.438 --> 33:58.859
[SPEAKER_02]: And this concludes another chapter of Coat Story.
34:08.294 --> 34:11.037
[SPEAKER_02]: code story is hosted and produced by no-alapart.
34:11.498 --> 34:15.943
[SPEAKER_02]: Be sure to subscribe on Apple podcast, Spotify, or the podcasting app at your choice.
34:16.904 --> 34:18.846
[SPEAKER_02]: And when you get a chance, leave us a review.
34:19.346 --> 34:21.389
[SPEAKER_02]: Both things help us out tremendously.
34:24.312 --> 34:25.413
[SPEAKER_02]: And thanks again for listening.